The many concepts, notions and tools that make up the digital world are constantly evolving, and are increasingly covered by the media, which makes them appear to be complex in several ways. This is the case of artificial intelligence, whose various and varied uses are becoming more and more common in our daily lives. These highly technical issues, often poorly understood by the general public, are the subject of extensive public debate. ACN believes that the level of trust people have in any technology, tool or digital application conditions its acceptance and therefore its use. That is why, when it comes to artificial intelligence, the companies in our industry decided very early on to participate in defining the objective criteria that constitute this level of confidence. This fundamental reflection leads to distinguish the different uses of AI and their associated risks in order to balance the defense of fundamental rights and freedoms, innovation and the development of European economic actors.
The emergence of French and European AI leaders, who develop their technologies while respecting our common fundamental values, is essential for the control of our future and our digital sovereignty. If in ten years the only tools available are designed, developed and controlled according to other value systems whose compatibility with our rights and freedoms must be questioned, then the primary objective of protecting these rights will be impossible to achieve.
Today, ACN maintains a privileged and constructive dialogue with public authorities and administrations in order to develop a common understanding of these issues. This is particularly the case with the CNIL regarding the deployment of “smart” cameras in public spaces, for which an ongoing conversation has enabled industry companies and the CNIL to share their respective visions, leading to changes in their positions.
Through this type of work, ACN wishes to contribute to the definition of what is a Trusted Artificial Intelligence. This reflection can be understood on many different but complementary levels.
First of all, the legal framework
Its role is to be both the guarantor of each citizen’s fundamental rights and freedoms, but also to allow the development of a strategic technology such as a trusted AI. The rules and guidelines governing AI systems, as well as the regulations and legislation, must ensure that the risks linked to the use of this technology are precisely delimited according to its uses (in accordance with what the European texts currently being drafted – the AI Act, AI Liability Directive – seem to foresee) and not ban the technology in itself, thus depriving us of tools that are essential to our digital sovereignty and strategic autonomy. In each of its initiatives, ACN highlights this approach to legislators, public authorities and administrations in order to help them build the necessary framework of trust and confidence in artificial intelligence.
Secondly, society’s understanding of the concept of digital trust
It is essential that this concept be sufficiently explicit and that the guarantees put in place, especially through the legal framework, be reassuring enough for society to focus on the benefits of the various AI technologies available. This requires an educational effort, an increase in awareness of the actual capabilities of AI, as well as debates around these technologies.
The French Human Rights Defender’s latest report on the “Perception of the development of biometric technologies in France” presents an overview of the state of knowledge and the way the French perceive biometric technologies. The study shows that the understanding of this technology is unclear and the degree of trust varies according to the entities that deploy it and the places in which it is used. However, citizens seem to be fully aware of the risks associated with these tools, which are perceived as being poorly regulated.
It is therefore important for ACN to make this reality known through a variety of projects, not only to institutions and decision-makers, but also to citizens, and to fight against the fantasy and demonization surrounding the topics and use of AI. For example, the modification of algorithmic settings is of such a technical complexity that only the intervention of an engineer specialized in this specific technology could have a notable effect modifying and/or diverting the function for which the skill was initially created. Settings that enable changing functions are not a simple configuration and depend on the construction of the algorithm that is not available to the operators on a daily basis.
Another way to restore trust in AI systems would be to introduce the technology through small-scale experiments, in the heart of everyday life, thus making the concepts, the tool and its settings more transparent.
Finally, to provide technological trust
A very important element to take into account is the level of trust in this technology itself, in the algorithms built in compliance with a national and European legal framework that protects against the abuses of AI and according to a principle of security and privacy by design. The citizens’ trust in AI technology will also reflect the ethical standards of the companies that develop and work with these systems, in their choice of algorithms and in their transparency, in their ability to reduce biases through improved learning, in the audits carried out to prevent false positives or false negatives, and so on.
Provided that the elements of trust are debated and shared, and that they lead to a legal framework and clear technical guidelines, AI will be able to flourish. More and more benefits are expected, such as in health, environment, flow management or security sectors. In today’s increasingly digitalized world, AI will be essential for us to manage the various challenges we face today, from geopolitical conflicts to the management of major events, including the energy transition and the improvement of daily life. Trust is at the heart of all these changes: this is why ACN wishes to continue to work alongside all the interested parties on the collective understanding and characterization of this level of trust and confidence, while at the same time contributing to the democratic reflections that are essential for the serene development of these solutions.